Model training method and device

The invention discloses a model training method and a model training device, which are applied to a machine learning platform. The machine learning platform comprises a plurality of functional components, wherein the functional components are used for implementing specified functions in a model training process. The model training method comprises the steps of: generating a directed functional component group in response to a group constructing instruction and based on a plurality of the selected functional components and an execution sequence of the plurality of functional components; receiving multiple groups of parameters set for the directed functional component group; and running the directed functional component group according to the multiple groups of parameters circularly in response to a running instruction. 说明书披露种模型训练方法和装置，应用于机器学习平台，所述机器学习平台包括若干功能组件，所述功能组件用于实现模型训练流程中的指定功能，所述模型训练方法包括：响应于建群指令，基于选定的多个功能组件和所述多个功能组件之间的执行顺序生成有向功能组件群；接收为所述有向功能组件群设置的多组参数；响应于运行指令，根据所述多组参数循环运行所述有向功能组件群。